Many organizations allow customers, business partners and others to communicate with them using electronic messages such as email. Some such organizations find themselves receiving large volumes of messages each day, and may therefore have implemented one of the currently available email response management systems. Such a system may handle 400,000 or more incoming customer email messages each day. These incoming email messages may relate to any number of different business scenarios or processes. For example, certain customers may submit service order requests to their service provider via email. Other customers may submit complaints or assistance requests to certain providers via email. If there is a very large volume of such received messages and there is great diversity in the nature of the individual messages, then such a system's usefulness will depend largely on how intelligently it is able to process the messages.
One approach is to provide automatic processing of the incoming email messages. In this approach, the system attempts to automatically analyze the content of the messages in a semantic way, perhaps using a rule-based engine in generating an acknowledgement or response message to send back to the customer. The implementation of such an approach may be quite difficult because a sophisticated linguistic analysis may be required, and because the complex knowledge management systems that implement such an approach may be difficult and expensive to maintain.
Another approach is to provide a certain level of automated processing of the email messages in combination with user intervention. The system may be able to partially analyze or categorize the incoming email messages, and the user may then make subsequent decisions on how to route or respond to the messages.
Existing systems using either of these approaches, however, may have disadvantages in their processing of received messages. Such systems may take into account only a limited scope of information relating to the processed message. Moreover, they may be restricted in the way they apply processing rules to the information they retrieve. That is, while existing systems may be able to gather information relating to a particular message and use that information in deciding how to process the message, they may not be flexible enough to provide a sufficient business context for processing the received message. Accordingly, these systems may yield unsatisfactory results in some situations.